Anomie behavior in sports competitions among college students remains a significant concern. Existing empirical studies have established a significant correlation between organizational factors and sports anomie behavior. However, traditional statistical methods have limitations in uncovering the intricate relationships and contributions of these factors to the occurrence of anomie behavior. This study explores the application of machine learning algorithms in predicting such behaviors, focusing on organizational factors, aiming to provide novel insights and analytical approaches. Method: A questionnaire was designed based on validated measurement scales. A total of 5,012 questionnaires were distributed to students from six universities in Xi'an, China, with 4,776 valid responses. The sample was balanced in terms of gender, academic year, and family background. After cleaning and preprocessing the data, 80% were used for training, and 20% for testing. The task of predicting sports anomie behavior was treated as a risk probability regression problem. A Random Forest regression model was built using Python and the Scikit-learn library. Hyperparameter optimization was performed to enhance model performance. The final model was selected after iterative evaluations, and feature importance analysis was conducted to assess the contributions of various organizational factors to anomie behavior. The proposed model performed well on the test data, with a Mean Squared Error of approximately 0.026, Mean Absolute Error of 0.12, and a coefficient of determination (R²) of 0.65. The Median Absolute Error was 0.09. Compared to traditional linear regression, the proposed model showed better fitting accuracy, with R ² improving by 0.27. Overall, its prediction accuracy reached 86%. Feature importance analysis showed that among organizational factors, "Rules and Regulations" has a relative importance of 27.5%. This was followed by "Behavioral Supervision" (20%), "Moral Education" (17.5%), "Promotion of Regulations" (15%), "Rewards and Punishments" (12.5%), and "Teaching and Training" (7.5%). Conclusion: This study demonstrates the feasibility of applying machine learning to predict and analyze sports anomie behavior. The proposed model reveals that "Rules and Regulations" is the most significant organizational factor influencing anomie behavior in college sports competitions, while "Teaching and Training" has a limited impact. The findings provide a new analytical perspective for researchers and offer data-driven evidence that can assist administrators in managing and intervening in anomie behavior in sports competitions.
Wang et al. (Thu,) studied this question.